Viderity Inc.’s Post

Our next Software Engineering Leaders Forum (SELF) will take place on 21st May! This event allows ScotlandIS members to consider a topic and share challenges and experiences in a safe space, with guest speaker Paul George from TAAP kicking off the session around the subject of Generative AI. If you're a CISO, CTO or software engineering leader, book your space now! https://lnkd.in/eScM48ff #software #engineering #generativeAI

Our next Software Engineering Leaders Forum (SELF) will take place on 21st May! This event allows ScotlandIS members to consider a topic and share challenges and experiences in a safe space, with guest speaker Paul George from TAAP kicking off the session around the subject of Generative AI. If you're a CISO, CTO or software engineering leader, book your space now! https://lnkd.in/eScM48ff #software #engineering #generativeAI

Today marks the day. I am now 2 years on LinkedIn 🎉🥳! Any tips or tricks you'd like to share? How can I make more content more interesting? What would you like to hear about? There are a lot more posts about architecture and scalability incoming in the next year. 🙏 Thanks to everyone for following and participating in the discussions!

Excellent review, Dorian. I love your clear explanation and analysis. Highly recommended. ⭐⭐ ⭐ ⭐ ⭐ #InformationArchitecture #SpecificityGradient

It’s exciting, but certainly not easy… With centralised data teams navigating the delicate balance between innovation, federation and maintaining governance. A C-Suite hungry for rapid return on investment from years of building out their data and cloud capabilities. An ever evolving technology ecosystem developing capability possibly quicker than many enterprise architects can stay abreast of. Predicting where the tech talent gaps will exist in 24 months rather than now is as much art as a science. To unlock advanced analytics and AI, data quality and robust data management is a non-negotiable… Can we deliver more with less? Read our full report on the challenges of reshaping tech strategy for free now: https://lnkd.in/d3SMq3T4

Since IAC - Information Architecture Conference is coming up I decided to review my talk from last year: https://lnkd.in/gqz2AxNr Summary: The Specificity Gradient is a way to think about how knowledge is built up to deliver products and services.

This video is about how LlamaIndex is building a new kind of knowledge assistant that is more powerful and flexible than current systems. 1. Challenges of current knowledge assistants: Current knowledge assistants are limited in their ability to understand complex queries and perform multiple tasks. They typically rely on a Retriever-Augmented Generation (RAG) pipeline, which is a simple search system on top of retrieval methods. 2. LlamaIndex proposes a three-step approach to address the limitations of current knowledge assistants: a. Advanced Data and Retrieval Modules: This involves using good data processing techniques, such as parsing and indexing, to improve the quality of data fed into the system. b. Advanced Single Agent Query Flows: This involves building on top of a basic RAG system with additional functionalities like query planning and agent reasoning loops. This allows the system to better understand and answer complex queries. c. General Multi-Agent Task Solvers: This involves using multiple agents, each specializing in a specific task, to work together to solve a complex problem. This approach is more scalable and efficient than using a single agent for everything. 3 Llama Agents is a framework for building multi-agent knowledge assistants. Llama Agents allows developers to create agents as microservices that can communicate with each other and work together to solve tasks. The video also includes a demo of how Llama Agents can be used to build a knowledge assistant on top of a basic RAG pipeline.

How may our industries inject and elevate graph and semantic standards under this effort? I know there are companies and businesses that would benefit with dogs for the fight, and provide leverage towards knowledge utilization across multiple uses. For instance, the USGS Geospatial Semantics and Ontology section is still working on their "Map as a Knowledge Base" program for richer topographic and geospatial data models. (https://lnkd.in/e8Y3S8iZ) This is within the Fed structure. Sponsors and champions are needed to collaborate from without to steer us and the world to a richer, more informed understanding of all the 'things' we claim to know.

Great summary of where #citizendevelopment is now and where it’s heading. It begs a question I’d love input on: the article talks about how GenAI will eliminate the need for many users to know any code at all. But I’m also seeing a sort-of doubling down on code in tools like Databricks’s AI assistant. You can ask it a question and it will write the code to get you to the answer - but the output is still code, which is hard for non-tech users to interpret, evaluate, debug, and enhance. Where do we think this will shake out? I expect to see UI layers come out on top of such assistants to make the code writing easier. There are plenty of drag-and-drop coding solutions out there but most people in tech don’t want to use them, for a variety of reasons. What does a UI layer look like that meets citizen needs AND tech needs? Will such a thing exist, or will there always be a divide between citizen needs and tech needs?

Are you a CISO, CTO or software engineering leader? Don't miss our next Software Engineering Leaders Forum (SELF) on 21st May! This event allows ScotlandIS members to consider a topic and share challenges and experiences in a safe space, with guest speaker Paul George from TAAP kicking off the session around the subject of Generative AI. Book your space now! https://lnkd.in/eScM48ff #software #engineering #generativeAI